Towards proteome database of Francisella tularensis

Journal of Chromatography B, 787 (2003) 149–177 www.elsevier.com / locate / chromb

Review

Towards proteome database of Francisella tularensis ´ a , Lenka Hernychova´ a , Jana Havlasova´ a , Irena Kasalova´ a , Martin Hubalek ˇ Neubauerova´ a , Jirı ˇ´ Stulık ´ a , *, Alesˇ Macela a , Margaretha Lundqvist b , Par ¨ Larsson b Vera a

ˇ ˇ ´ 1575, Proteome Center for the Study of Intracellular Parasitism of Bacteria, Purkyneˇ Military Medical Academy, Trebesska ´ ´ , Czech Republic 500 01 Hradec Kralove b Swedish Defence Research Agency, SE 901 82, Umea˚ , Sweden

Abstract The accessibility of the partial genome sequence of Francisella tularensis strain Schu 4 was the starting point for a comprehensive proteome analysis of the intracellular pathogen F. tularensis. The main goal of this study is identification of protein candidates of value for the development of diagnostics, therapeutics and vaccines. In this review, the current status of 2-DE F. tularensis database building, approaches used for identification of biologically important subsets of F. tularensis proteins, and functional and topological assignments of identified proteins using various prediction programs and database homology searches are presented.  2002 Elsevier Science B.V. All rights reserved. Keywords: Review; Proteome databases; Francisella tularensis

Contents 1. Introduction ............................................................................................................................................................................ 1.1. Outline of the current status of bacterial proteome studies .................................................................................................. 1.2. Biological and clinical importance of F. tularensis ............................................................................................... 2. Proteome analysis of F. tularensis LVS .................................................................................................................................... 2.1. Construction of reference protein maps of whole cell lysates .............................................................................................. 2.2. Mapping of integral membrane proteins............................................................................................................................ 2.3. Immunoreactive proteins of F. tularensis LVS microbes .................................................................................................... 2.4. F. tularensis protein profiling under conditions mimicking a hostile intracellular environment.............................................. 3. Comparative analysis of F. tularensis type A and B strains ........................................................................................................ 4. Construction of virtual and real proteome of F. tularensis—from genome to proteome ................................................................ 5. Conclusions ............................................................................................................................................................................ Acknowledgements ...................................................................................................................................................................... References ..................................................................................................................................................................................

149 150 152 153 153 164 165 167 171 173 174 175 175

1. Introduction *Corresponding author. Fax: 1420-49-551-3018. ´ E-mail address: [email protected] (J. Stulık).

Despite advanced antibiotic therapy, infectious and parasitic diseases still represent the main cause

1570-0232 / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S1570-0232(02)00730-4

150

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

of death for mankind. The reality can be even worse due to the increasing number of bacterial strains exhibiting resistance against antibiotics on the one hand and the increased resurgence of seemingly eliminated infections such as tuberculosis or cholera on the other. Thus the only reaction to this new situation should be diligent effort aimed at identifying new molecules with diagnostic and vaccine applications and determining novel targets for more efficient drug therapy. This challenging task is partly alleviated by microbes themselves. Microbes are relatively simple entities that can be grown in chemically defined media providing scientists with sufficient amounts of material for analyses. Furthermore, bacterial genomes are small, hence current technologies, in spite of their inherent limitations, can be successfully applied for the comprehensive scrutiny of microbial genome expression. The restricted scope of prokaryotic genomes is also responsible for the fact that over 90% of the completely sequenced genomes are of microbial origin. According to an updated list (see www.tigr.org) more than 80 microbial genomes have been completely sequenced and, additionally, more than 130 microbial genome-sequencing projects are in progress. The availability of complete genome sequences is prerequisite for more targeted approaches focused on detection of molecules of significance. Until recently, the approaches performed at DNA or RNA level, like comparative genomics or mRNA microarray technology, predominantly throve on genome knowledge. However, both approaches are defective in predicting corresponding protein expression level, its structural modification, cellular localization and possible protein–protein interactions. Therefore, simultaneous analysis of global protein patterns must complement ongoing genetic studies in order to provide more valid information for definitive selection of diagnostic or therapeutic candidates. The rapidly developing procedures combining two-dimensional gel electrophoresis (2-DE) or multidimensional chromatography with mass spectrometry (MS) techniques [1,2], summarized by the term proteomics, are expected to bring major breakthroughs in protein mapping. This review highlights some recent achievements in the application of proteomics for the analysis of bacterial proteomes. Special emphasis is placed on building a proteome

database of the bacterial intracellular pathogen Francisella tularensis.

1.1. Outline of the current status of bacterial proteome studies The current studies of bacterial proteomes can be basically divided into two large groups. The first one is concentrated on the building of comprehensive 2-DE protein databases containing the reference 2DE maps of individual organisms (see http: / / us.expasy.org / ch2d / 2d-index.html). The formation of these databases is now facilitated by concomitant progress in 2-DE, protein staining and, especially, in MS procedure. These databases should serve as the basis to which all other experimental strategies utilizing virulent, mutant strains or strains cultivated under harsh conditions can be then compared. However, all displayed databases exhibit one dominant shortcoming referring to the disproportion between the number of identified genes and the actual number of identified proteins. The reason for this imbalance is the failure of current proteomics technology to provide adequate resolution of proteins with extreme physicochemical properties and, further, the insufficient detection of proteins occurring in cell only in a few copies [3]. Most attention regarding microbial 2-DE database construction is paid to microorganisms that are clinically relevant and / or that have been extensively studied with respect to their genetics and biochemistry for many years. Furthermore, the accessibility of complete genome sequences is also crucial. Twodimensional protein databases of Escherichia coli and Haemophilus influenzae belong to the largest ones. As for the former, the recent application of ultra narrow pH gradients enabled display of more than 70% of the entire E. coli genome [4]. As for the latter the commonly applied proteomics procedures were extended to the application of several chromatographic steps, including heparin chromatography, chromatofocusing and hydrophobic chromatography. In this way 502 different proteins were identified in what represents about one third of completely sequenced open reading frames (ORF) [5]. The other intensively studied human pathogens are Mycobacterium tuberculosis [6] and Helicobacter pylori [7].

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

The alternative approach that should expand the number of identified proteins in 2-DE databases is based on the proteome analysis of purified protein subsets. This is particularly important when membrane and secreted proteins that significantly contribute to the mechanism of pathogenicity have to be studied. Triton X-114 or carbonate extraction followed by solubilization of purified proteins in buffers containing new types of detergents are methods of choice for isolation of integral membrane proteins [8,9]. These protocols were successfully applied in the global analysis of the outer membrane proteome of Leptospira interrogans and Caulobacter crescentus [10,11]. Bacterial secreted proteins play very diverse roles in host–pathogen interactions. They can reorganize host cytoskeletal structures, modulate cell-signaling pathways and protect engulfed bacteria against toxic molecules produced by infected cell [12]. Additionally, some of the secreted antigens exert strong immunostimulating effects. Weldingh et al. [13] established 2-DE reference maps of M. tuberculosis culture filtrate proteins in order to identify candidate antigens for a novel subunit vaccine against tuberculosis. In the case of Bacillus subtilis the release of extracellular proteins is associated with the generation of a heat-resistant endospore that enables bacteria to survive under poor nutrient conditions. Examination of the proteome of B. subtilis extracellular proteins led to the visualization of more than 100 spots. Of them, over 90% disappeared when mutant strains with deficient Sec protein-secretion pathways were tested [14]. The second group of proteome studies is aimed at identification of proteins whose expression relates to the pathogenicity of wild bacterial strains. The simplest approach exploits the comparative proteome analysis of protein patterns of non-virulent strains and their pathogenic counterparts. This procedure was successfully applied in comparative analysis of protein profiles of avirulent vaccine and virulent laboratory mycobacterial strains where besides the complex cell lysates the cell culture supernatants were also used for strain comparison. The results confirmed the existence of differences in both intensity and mobility between cell proteins extracted from mycobacterial strains differing in pathogenicity [15]. Similarly, unique proteins were found in protein spectra of genetically indistinguishable Pseudo-

151

monas aerigunosa strains representing initial and chronic isolates from a cystic fibrosis patient [16]. The more sophisticated approach to highlight proteins potentially associated with the progression of bacterial infection is the proteome study of bacteria cultivated under conditions mimicking hostile intracellular milieu or even after their ingestion by phagocytes. Regarding environmental influences the bacterial responses to heat, oxidative, acid stress and to nutrition defects were studied on the proteome level. Each stress condition induced its own distinctive set of genes, which partly overlapped, especially, in the overproduction of chaperonins [17–19]. Likewise, heat-shock proteins were up regulated in Brucella abortus, Leishmania chagasi and Legionella pneumophila growing inside infected host cells [20–22]. As for L. pneumophila the induction of cpn60 occurred very early in the course of infection and it was characteristic only for virulent strains. The highly decisive proteome study from the point of view of vaccine construction is the mapping of bacterial immunorelevant antigens. 2-DE immunoblotting utilizing sera collected from patients as primary antibodies is the predominant technique for monitoring candidate antigens. Haas et al. [23] performed an extensive immunoproteomics study of H. pylori infection in which the potential association between specific immune response and manifestation of disease was investigated. They compared the immunoreactivity of serum antibodies collected from patients suffering from active H. pylori infection, with a control group with unrelated gastric diseases and, finally, with patients with gastric cancer. The preliminary results confirmed the existence of antigens differently recognized by sera from gastritis and ulcer patients suggesting them as possible indicators of clinical manifestation [23]. Proteomics can also be a powerful method for elucidation of immunodominant T cell antigens. Using two-dimensional liquid phase electrophoresis, Covert et al. [24] separated M. tuberculosis culture filtrate and cytosolic proteins into great number of fractions which were tested for T cell stimulating activity via production of interferon g. Proteins occurring in positively tested fractions were then identified by liquid chromatography-mass spectrometry. Globally, of the 30-mycobacterial proteins with T cell stimulating activity identified, 17 of them were novel antigens.

152

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

1.2. Biological and clinical importance of F. tularensis F. tularensis, the causative agent of tularemia, was first observed in animal tissues by McCoy [25], subsequently isolated by McCoy and Chapin from ground squirrels in Tulare County, California, was considered a ‘‘plague-like disease’’, and named Bacterium tularense [26]. Shortly thereafter, tularemia was recognized as a rare but potentially severe and fatal illness in humans [27]. The disease is endemically spread over the northern hemisphere. Various small mammals, mainly members of the orders Rodentia and Lagomorpha, create the principal reservoirs in nature. Transmission to humans and other vertebrates can be mediated by bites of ectoparasites such as ticks, mites or deer flies, by handling or ingestion of infected material and / or water, and by inhalation of contaminated dust particles. The microbes are small, non-motile, nonsporulating gram-negative coccobacilli, which are nutritionally fastidious and require cysteine or Nathioglycolate as reducing agent for their growth in vitro [28,29]. According to its lifestyle, F. tularensis is a facultative intracellular bacterium proliferating in monocyte-macrophage cells and hepatocytes. Originally, two main types (genotypes, biovars) of F. tularensis were distinguished: F. tularensis tularensis occurring in the New World, and F. tularensis palaearctica (holarctica) spread over whole northern hemisphere, tentatively also designated as type ‘‘A’’ and ‘‘B’’, respectively [30,31]. In addition two other biovars (subspecies), F. tularensis palaearctica subvar japonica and F. tularensis biovar mediaasiatica were proposed later on [32]. The intraspecies position of last two sub-variants of F. tularensis remains to be definitely specified. F. tularensis mediaasiatica seems to be included in subsp. tularensis (type A) and subvar japonica form rather separate subspecies [33,34]. In 1940s, Gajskij repeatedly cultured the F. tularensis biovar holarctica on artificial media supplemented with antiserum [35]. Of several attenuated strains, such as Moscow, Ondatra IV, 155, and 10, Elbert consequently used strain No. 15 for construction of live vaccine against tularemia (for review, see Ref. [36]). After re-isolation of individual colonies derived from vaccine produced in the Gamaleia

Institute, Eigelsbach and Downs gave rise the live vaccine strain designated LVS in 1950s [37] (for review, see Ref. [38]). This LVS strain was used for vaccination in the USSR and Eastern Europe, and for protection of laboratory workers at USAMRIID, Fort Detrick, Frederick, MD, USA. The data presented in research reports from the USSR and Czechoslovak laboratories and retrospective study of laboratoryacquired tularemia done in the 1970s documented the partial protective effect of live vaccines against virulent tularemic strains [39–41]. Both strains Gajskij 15 and LVS are infectious for mice eliciting disease, which is in its main characteristics similar to human tularemia. The experimental murine tularemia induced by attenuated strains has been broadly used as the model for the study of pathogenesis of tularemia and studies of innate and adaptive immune mechanisms which are prerequisite for the expression of the protective effect of vaccination and natural infection against subsequent challenge [42,43]. Human tularemia is a sub-acute, usually moderately severe anthropo-zoonotic disease. The clinical manifestation of the infection seems to be dependent on the route of transmission, the genetic background of the host, and the virulence of the infecting microorganism. Ulceroglandular or glandular, oculoglandular, oropharyngeal, typhoidal, and respiratory (pneumonic) tularemia are distinct forms of tularemia according to the portal of entry. In general, infected patients expressing arbitrary form of tularemia may have rapid onset of fever accompanied with cough, creation of granulomas and in the case of more severe disease, with secondary pneumonia, which might be attributed to a transient bacteriemia. Mortality rates range from less than 1% in Eurasia, where the subspecies F. tularensis holarctica with inherent lower virulence are endemic, to 30% in North America, where the subspecies F. tularensis tularensis causes the most severe pulmonary disease. F. tularensis should accomplish several essential events to overcome the structural barriers lying between the microbe and the intracellular niche suitable for its proliferation. Microbes must first adher to the target cell followed by entering the host cells by the process called induced phagocytosis. The precise molecular base of this process is currently

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

unknown. Once internalized, F. tularensis microbes proliferate inside the cells. Likewise the process of cell entering, the compartment, where the microbes undergo intracellular multiplication has been poorly defined until now. Nevertheless, at least in the early stages of bacteria–macrophage interaction, the compartment can be characterized as an acidic vesicle that facilitates high input of iron essential for the growth of F. tularensis microbes [44]. In in vitro macrophage model system, intracellular multiplication of F. tularensis resulted in cytopathogenic effect [45] accompanied by microbe-induced apoptosis of infected cultured cells [46]. The induced programmed destruction of infected cell can ensure the transit of bacterium from a disrupted ‘‘exhausted’’ cell to a different one, originally an uninfected cell. The mechanism by which the tularemia microbe induces the fatal end of an infected host cell remains obscure because, unlike many other intracellular bacteria, F. tularensis has no toxic LPS and does not produce any known toxin. Macrophages, which are targeted by F. tularensis microbes, were identified to also be key cells in the process of resolution of infection. The elimination of bacteria from tissues of an infected host is highly dependent on the mechanisms of innate as well as adaptive immunity, both under multigenic control. The immunoregulatory pathway of Th1 type characterized by IFN-g, TNF-a and IL-2 production is preferentially activated in the course of primary infections of experimental animal models and after vaccination of human beings with live vaccine strain [47,48].

2. Proteome analysis of F. tularensis LVS

2.1. Construction of reference protein maps of whole cell lysates As mentioned above, the formation of comprehensive 2-DE microbial databases is prerequisite for subsequent comparative studies of vaccine and virulent strain protein profiles, for the analysis of the phenotype of multigenic responses invoked, for example, by cultivation of microbes under stressful conditions and, finally, for confirmation of the translation of the predicted ORFs. The construction

153

of 2-DE F. tularensis reference maps is based mostly on the analysis of total cellular proteins extracted from vaccine strain F. tularensis LVS. The solubilized proteins were separated on 2-DE gels of two different immobilized pH gradients (IPG). The second dimension separation on gradient 9–16% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) or 16.5% T, 6% C tricine SDS–PAGE then allowed resolution of proteins with molecular weights (MW) ranging from 8 to 200 kDa and 5 to 49 kDa, respectively. To increase the number of proteins entering the gels, the cell extracts were dissolved in isoelectric focusing (IEF) buffer containing thiourea and tributylphosphine which improve protein solubilization for 2-DE [49], and 150 mg of protein was loaded overnight by in-gel rehydration [50]. Using a broad range pH 3–10 IPG strips, more than 1500 spots were detected (Fig. 1). Application of the wide pH range gradient usually leads to optimal resolution of proteins with pI values from 4.0 to 7.5; however, this gradient fails to perform adequate separation of more basic spots. Therefore, an additional overlapping gradient of pH 6–11 was utilized for construction of 2-DE reference map of basic tularemic proteins. As can be seen from Fig. 2, the application of basic pH gradient did not significantly increase the number of detected basic spots in comparison to the wide range pH gradient, nevertheless the basic proteins were much better resolved because they were allowed to reach positions corresponding to their pI values. Using basic range IPG strip of pH 6–11, the separation of |500 spots was repeatedly accomplished. The application of tricine SDS–PAGE [51] in the second dimension was aimed improving the resolution of low molecular weight proteins normally hidden in the bottom part of the gel containing the mixture of ampholytes and detergents. The reference tricine SDS–PAGE gel containing |1000 spots is shown in Fig. 3. Compared to classical gradient 9–16% SDS–PAGE (Fig. 1), tricine SDS–PAGE gel provides very good resolution of proteins with molecular weights from 3 to 30 kDa. Recently, we got access to complete cell lysates of F. tularensis type A highly virulent strains. Availability of this material enabled comparative proteomic studies of type A and type B strains (see below) to be started and a reference 2-DE map of this most

154

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

Fig. 1. Silver-stained 2-DE reference map of F. tularensis LVS. Extracted proteins were resolved by IEF in the pH range 3–10 followed by SDS–PAGE gradient gel (9–16%). The spot numbers indicate identified proteins.

virulent strain to be constructed. The reference map of F. tularensis type A strain is shown in Fig. 4. This wild type strain, designated 176, was isolated from a blood sample of a patient with tularemia and belongs to subspecies F. tularensis tularensis.

The proteins indicated in all 2-DE reference maps by their spot numbers were identified by peptide mass fingerprinting. When the genome sequence of F. tularensis was not accessible the obtained mass spectra were matched against SWISS-PROT or

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

155

Fig. 2. Silver-stained 2-DE reference map of F. tularensis LVS. Extracted proteins were resolved by IEF in the pH range 6–11 followed by SDS–PAGE gradient gel (9–16%). The spot numbers indicate identified proteins.

NCBI protein databases where amino acid sequences of |80 tularemic proteins are located. This approach was not very successful, resulting in unambiguous identification of only 19 different protein species. All data acquired from this search are summarized in

Table 1. Most identified proteins belong to the molecular chaperone system of F. tularensis. It is interesting to note that several charge and mass variants of the 10-kDa chaperonin were identified on 2-DE maps of both vaccine and highly virulent strain

156

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

Fig. 3. Silver-stained 2-DE reference map of F. tularensis LVS. Extracted proteins were resolved by IEF in the pH range 3–10 followed by tricine SDS–PAGE gel (16.5% T, 6% C). The spot numbers indicate identified proteins.

176. While the MW/ pI 9.7 / 4.9 10-kDa chaperonin variant was common for both types of strains, the distribution of the others in vaccine and virulent strain can be clearly distinguished. At the end of 2001 the annotation of partial genome sequence of F. tularensis strain Schu 4 was published [52]. In total

1804 candidate ORFs were identified in the data set of which 1289 were thought to encode proteins (743 genes with known function, 133 genes coding for hypothetical proteins and 413 genes with no database match). Currently, the protein database created by transla-

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

157

Fig. 4. Silver-stained 2-DE reference map of wild type F. tularensis strain, designated 176. Extracted proteins were resolved by IEF in the pH range 3–10 followed by SDS–PAGE gradient gel (9–16%). The spot numbers indicate identified proteins.

tion of all possible ORFs is available for matching process. The amino acid sequences obtained from the ORF database are utilized in the search for proteins with sequence homology of other microbial species listed in the NCBI database using the BLAST

program. By this approach 124 proteins have so far been newly annotated. The basic characteristics of identified proteins together with the source of their homologues are listed in Table 2. Among identified spots only one protein (spot no. 943, tricine SDS–

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

158

Table 1 The list of identified proteins of F. tularensis included in the NCBI and SWISSPROT databases Spot No.

MW (kDa) / pI, measured

Sequence coverage, %

120 214 790 793 920 931 939 962

Chaperone protein dnaK 60-kDa Chaperonin Hypothetical 23-kDa protein Hypothetical 23-kDa protein 10-kDa Chaperonin 10-kDa Chaperonin 10-kDa Chaperonin 10-kDa Chaperonin

Whole-cell antigen F. tularensis LVS IEF pH 3–10 9–16% SDS–PAGE P48205 69.2 / 4.9 P94798 57.4 / 5.0 CAA70085 22.1 / 5.7 CAA70085 22.1 / 5.7 P94797 10.3 / 5.5 P94797 10.3 / 5.5 P94797 10.3 / 5.5 P94797 10.3 / 5.5

69.2 / 4.9 55.6 / 4.9 23.3 / 5.4 23.3 / 5.2 10.3 / 5.2 9.7 / 4.9 9.3 / 5.3 8.6 / 5.3

13 30 40 24 32 20 38 20

488 495

Whole-cell antigen F. tularensis LVS IEF pH 6–11 9–16% SDS–PAGE Putative peptidyl-prolyl cis-trans isomerase AAG33118 10.2 / 6.5 Putative peptidyl-prolyl cis-trans isomerase AAG33118 10.2 / 6.5

14.6 / 5.3 14.5 / 6.1

64 58

115 204 792 903

Chaperone protein dnaK 60-kDa Chaperonin Hypothetical 23-kDa protein Macrophage growth locusB

Whole-cell antigen F. tularensis LVS IEF pH 3–10 16.5% T, 6% C tricine SDS–PAGE P48205 69.2 / 4.9 P94798 57.4 / 5.0 CAA70085 22.1 / 5.7 AAC29033 15.1 / 4.1

69.2 / 4.9 60.3 / 4.9 23.0 / 5.5 7.8 / 4.7

26 37 18 16

47 236 275

Integral membrane proteins F. tularensis LVS IEF pH 3–10 9–16% SDS–PAGE 60-kDa Chaperonin P94798 57.4 / 5.0 Hypothetical 23-kDa protein CAA70085 22.1 / 5.7 17-kDa Major membrane protein (precursor) P18149 16.0 / 9.2

57.4 / 5.0 23.0 / 5.4 13.3 / 4.8

14 34 56

Whole-cell antigen F. tularensis subsp. tularensis, strain 176 IEF pH 3–10 9–16% SDS–PAGE Chaperone protein dnaK P48205 69.2 / 4.9 Acid phosphatase AAB06624 57.6 / 5.9 Cell division protein FtsZ AAC99558 39.7 / 4.8 Triosephosphate isomerase P96763 27.6 / 4.9 GrpE protein P48204 22.0 / 4.8 GrpE protein P48204 22.0 / 4.8 Hypothetical 23-kDa protein CAA70085 22.1 / 5.7 Hypothetical 23-kDa protein CAA70085 22.1 / 5.7 Hypothetical 23-kDa protein CAA70085 22.1 / 5.7 10-kDa Chaperonin P94797 10.3 / 5.5 10-kDa Chaperonin P94797 10.3 / 5.5 10-kDa Chaperonin P94797 10.3 / 5.5

69.2 / 4.8 59.0 / 6.0 46.9 / 4.7 27.4 / 4.9 26.6 / 4.8 26.6 / 4.8 23.0 / 5.8 23.0 / 6.0 23.0 / 5.5 10.0 / 4.8 9.7 / 4.9 9.7 / 5.9

28 23 26 24 40 26 45 30 24 20 21 37

390 673 1149 1745 1761 1768 1843 1854 1863 2105 2121 2127

Protein name

Accession No.

MW (kDa) / pI, theoretical

The mass spectra were recorded in reflector mode on a MALDI mass spectrometer Voyager-DE STR (Perseptive Biosystems, Framingham, MA, USA) equipped with delayed extraction. Proteins were identified by peptide mass fingerprinting using ProFound and PeptIdent programs. The spectra of integral membrane proteins were recorded on a LC–MS–MS mass spectrometer Q-TOF Ultima姠 API (Micromass, UK) fitted with a Z-spray atmospheric pressure ionization (API). The in-gel protein digest samples were separated by means of a Micromass modular CapLC system (Micromass, UK) connected directly to the Z-spray source. The MS–MS spectra were acquired on the four most intense ions. All data were processed automatically by means of ProteinLynx software. Protein identification was done by searching the Non Redundant Data Base (NRDB) using the ProteinLynx Global Server engine.

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

159

Table 2 The list of identified proteins of F. tularensis-homologues according to BLAST Spot No.

MW (kDa)/pI, theoretical

MW (kDa)/pI, measured

Sequence coverage, %

340

41.8/5.0

47.8/4.6

30

488

39.6/5.7

40.3/5.9

33

753

26.0/6.1

25.7/6.3

22

755

25.7/7.7

25.6/5.3

23

817

23.3/5.2

20.3/4.7

23

858

18.2/6.8

15.3/6.8

32

877

18.8/7.7

14.4/6.4

34

889

16.5/5.0

13.7/4.9

42

893

16.1/5.6

13.3/5.4

45

918

18.5/5.3

10.3/5.4

40

933

12.8/4.6

9.6/4.6

20

949

16.3/6.1

9.4/6.0

42

982

11.2/5.9

7.8/6.3

46

49

52.1/7.2

52.9/7.9

33

56

52.1/7.2

52.5/6.7

27

72

49.2/6.5

50.1/5.4

32

77

49.2/6.5

49.2/6.0

32

91

49.2/6.5

48.5/6.6

32

110

48.1/8.6

46.2/8.6

12

Protein homologues according to BLAST Protein name Organism

Whole-cell antigen F. tularensis LVS IEF pH 3–10 9–16% SDS–PAGE Elongation factor TU S. typhimurium Glycine-cleavage system protein T1 Pseudomonas aeruginosa Oxidoreductase Vibrio cholerae Hypothetical protein Plasmodium falciparum Peptidoglycan-associated lipoprotein Pal Yersinia pestis (3R)-hydroxymyristoyl (acyl carrier protein) dehydratase P. aeruginosa Cationic 19-kDa outer membrane protein precursor Yersinia enterocolitica Biotin carboxyl carrier protein P. aeruginosa 50S Ribosomal protein L9 P. aeruginosa Probable bacterioferritin P. aeruginosa 50S Ribosomal protein Campylobacter jejuni 3-Dehydroquinase Buchnera sp. Probable sigma (54) modulation protein Pseudomonas putida Whole-cell antigen F. tularensis LVS IEF pH 6–11 9–16% SDS–PAGE Inosine-59-monophosphate dehydrogenase P. aeruginosa Inosine-59-monophosphate dehydrogenase P. aeruginosa Glutamate dehydrogenase Clostridium symbiosum Glutamate dehydrogenase C. symbiosum Glutamate dehydrogenase C. symbiosum Penicillin-binding protein 5 P. aeruginosa

E value Accession No.

P21694

1310 2168

G82994

2310 297

G82383

2310 284

CAB38995

0.12

CAC89968

3310 220

AAG07033

3310 231

P31519

2310 211

P37799

6310 227

F83029

2310 230

B83036

3310 24

H81392

3310 225

P57479

5310 230

P15592

2310 223

H83173

1310 2172

H83173

1310 2172

P24295

1310 2164

P24295

1310 2164

P24295

1310 2164

C83146

1310 265

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

160 Table 2. Continued Spot No.

MW (kDa)/pI, theoretical

MW (kDa)/pI, measured

Sequence coverage, %

Protein homologues according to BLAST Protein name Organism

Accession No.

E value

161

39.9/8.5

37.9/8.6

16

G71856

5310 25

178

36.1/6.4

36.0/5.8

27

AAG41996

1310 2108

180

36.1/6.4

36.0/7

31

AAG41996

1310 2108

193

35.5/7.7

34.4/8.4

22

AAK02376

1310 2101

198

33.7/9.1

34.3/8.8

15

P36659

4310 234

217

24.6/9.5

32.1/9.0

17

P02384

4310 277

219

32.0/7.7

32.2/7.9

35

AAA93461

2310 245

220

24.6/9.5

31.9/8.9

24

P02384

4310 277

233

27.9/6.2

31.7/5.6

40

F83319

4310 211

256

26.6/7.9

30.0/8.7

22

CAA74088

4310 295

257

27.7/8.9

30.0/8.9

52

T43699

0.017

258

32.0/9.0

30.0/8.8

11

AAC43386

7310 262

259

25.5/6

29.7/5.0

38

F83048

8310 243

272

30.3/6.6

28.9/9.0

20

AAK58507

6310 273

286

27.7/8.9

27.6/8.6

50

T43699

0.017

291

23.4/9.3

27.1/8.9

23

P44920

2310 222

299

23.1/9.5

26.4/9.0

38

E64092

5310 274

305

24.4/9.0

26.0/8.8

22

BAB43788

1310 246

308

24.4/9.0

25.9/8.7

17

BAB43788

1310 246

331

21.0/8.5

23.1/8.8

21

AJ414148

4310 218

352

21.0/8.5

22.4/8.7

35

AJ414148

4310 218

362

20.0/8.7

21.4/8.8

15

AAG07663

2310 251

396

19.6/9.4

18.3/8.9

28

C82439

3310 254

400

15.8/6.2

18.1/6.5

21

P37812

1310 214

411

14.4/9.4

17.3/8.9

41

Hypothetical protein jhp1045 H. pylori Malate dehydrogenase Sinorhizobium melioti Malate dehydrogenase S. melioti Acetyl-coenzyme A carboxylase Pasteurella multocida Curved DNA-binding protein cbpA E. coli 50S Ribosomal subunit protein L1 E. coli IMI-1 Enterobacter cloacae 50S Ribosomal subunit protein L1 E. coli Probable thiosulfate sulfurtransferase P. aeruginosa Succinate dehydrogenase putative iron sulphur subunit Shewanella frigidimarina DNA mismatch repair protein msh2 Schizosaccharomyces pombe LPSA protein Dichelobacter nodosus Probable two-component response regulator P. aeruginosa L-Aspartate beta-decarboxylase Alcaligenes faecalis subsp. faecalis DNA mismatch repair protein msh2 S. pombe Dephospho-coenzyme A kinase H. influenzae 50S Ribosomal protein L3 H. influenzae (strain Rd KW20) Pyrrolidone-carboxylate peptidase Staphylococcus aureus subsp. aureus Mu50 Pyrrolidone-carboxylate peptidase S. aureus subsp. aureus Mu50 Conserved hypothetical protein Y. pestis Conserved hypothetical protein Y. pestis Transcription antitermination protein NusG P. aeruginosa Peptide methionine sulfoxide reductase V. cholerae ATP synthase Bacilus subtilis RpS8 P. multocida

AAK03485

6310 237

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

161

Table 2. Continued Spot No.

MW (kDa)/pI, theoretical

MW (kDa)/pI, measured

Sequence coverage, %

Protein homologues according to BLAST Protein name Organism

Accession No.

425

16.3/6.1

17.2/4.9

50

P57479

5310 230

437

14.0/9.3

16.7/8.7

18

H82154

5310 26

449

11.9/9.8

15.7/9.1

31

G82059

3310 245

450

16.1/6.6

15.7/6.6

36

T24480

0.019

466

11.2/5.9

15.4/5.4

60

T01754

2310 223

469

10.8/9.5

15.3/8.9

19

D82532

5310 222

476

12.2/7.7

14.8/8.4

37

P37395

1310 222

478

9.5/9.8

14.5/9.1

35

3-Dehydroquinase Buchnera sp. Conserved hypothetical protein VC1812 V. cholerae Ribosomal protein S10 V. cholerae Hypothetical protein T0H1.4 Caenorhabditis elegans Hypothetical protein 102 P. putida 50S Ribosomal protein L25 Xylella fastidiosa Thioredoxin Cyanidium caldarium Histone-like protein HU form B P. putida

AAK52475

3310 227

B82079

1310 2170

O51312

1310 2153

P21694

1310 2168

O87796

1310 2169

P57983

6310 277

CAB38995

0.12

P09157

1310 273

CAC47046

1310 265

F83029

3310 230

B83036

3310 24

–

–

NP 462604 ]

1310 2171

313

50.5/5.6

55.4/5.8

336

49.5/4.7

54.3/4.8

380

41.8/5.0

50.4/5.0

439

38.1/5.3

47.3/5.5

618

31.0/5.6

36.7/5.4

749

25.7/7.7

26.2/5.3

817

21.9/5.4

21.9/5.3

826

19.7/5.1

20.5/5.1

868

16.1/5.6

12.2/5.4

885

18.5/5.3

10.3/5.4

943

8.6/4.5

3.6/5.0

678

57.6/5.8

58.7/5.7

Whole-cell antigen F. tularensis LVS IEF pH 3–10 16.5% T, 6% C tricine SDS–PAGE 12 Pyruvate dehydrogenase V. cholerae 14 Enolase Borrelia burgdorferi 19 Elongation factor Tu S. typhimurium 14 Fructose-bisphosphate aldolase Pseudomonas stutzeri 22 Elongation factor TS P. multocida 38 Hypothetical protein P. falciparum 29 Superoxide dismutase E. coli 18 Conserved hypothetical protein Sinorhizobium meliloti 13 50S Ribosomal protein L9 P. aeruginosa 44 Probable bacterioferritin P. aeruginosa 59 Unknown protein Whole-cell antigen F. tularensis subsp. tularensis, strain 176 IEF pH 3–10 9–16% SDS–PAGE 19 Phosphoglyceromutase S. typhimurium LT2

E value

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

162 Table 2. Continued Spot No.

MW (kDa)/pI, theoretical

MW (kDa)/pI, measured

Sequence coverage, %

Protein homologues according to BLAST Protein name Organism

Accession No.

E value

686

57.6/5.8

58.7/5.8

19

NP 462604 ]

1310 2171

702

57.6/5.8

58.6/5.6

19

NP 462604 ]

1310 2171

898

48.6/6.1

52.8/6.0

16

AAG55380

1310 2143

916

49.1/6.5

52.4/6.4

12

P24295

1310 2161

919

49.1/6.5

52.4/6.1

22

P24295

1310 2164

994

50.3/6.4

50.2/6.5

13

NP 288589 ]

1310 2153

1007

50.5/5.6

50.2/6.3

11

NP 232042 ]

1310 2170

1091

39.7/6.0

48.4/6.0

11

NP 486532 ]

1310 2159

1096

47.0/6.2

48.5/5.9

13

NP 250271 ]

1310 2120

1101

47.0/6.2

48.5/6.1

18

NP 250271 ]

1310 2120

1116

47.0/6.2

48.4/5.9

13

NP 250271 ]

1310 2120

1233

38.2/5.3

44.7/5.4

24

O87796

1310 2169

1241

38.2/5.3

43.9/5.3

20

O87796

1310 2169

1278

41.9/5.5

42.9/5.2

13

NP 417401 ]

1310 2127

1287

38.2/5.3

42.6/5.3

20

O87796

1310 2169

1297

41.0/5.7

41.6/5.7

12

NP 521198 ]

3310 239

1303

39.0/5.4

41.5/5.2

30

NP 602691 ]

3310 218

1348

38.3/5.3

39.7/5.2

22

Q42689

4310 297

1354

38.3/5.3

40.0/5.1

30

Q42689

4310 297

1389

37.8/5.9

38.6/5.9

13

NP 391765 ]

1310 2116

1404

37.6/6.3

38.4/5.4

39

NP 562220 ]

1310 2129

1413

38.3/5.3

37.8/5.5

11

NP 230388 ]

2310 290

1423

36.1/6.4

37.6/5.9

22

NP 533304 ]

1310 2109

1435

30.1/6.1

36.1/5.9

21

Phosphoglyceromutase S. typhimurium LT2 Phosphoglyceromutase S. typhimurium LT2 Serine tRNA synthetase E. coli NAD-specific glutamate dehydrogenase C. symbiosum NAD-specific glutamate dehydrogenase C. symbiosum orf, hypothetical protein E. coli O157:H7 EDL933 Pyruvate dehydrogenase, E3 component, lipoamide dehydrogenase V. cholerae Probable ABC transporter membrane protein (ycf24) Nostoc sp. PCC 7120 Citrate synthase P. aeruginosa Citrate synthase P. aeruginosa Citrate synthase P. aeruginosa Fructose-bisphosphate aldolase P. stutzeri Fructose-bisphosphate aldolase P. stutzeri Phosphoglycerate kinase E. coli K12 Fructose-bisphosphate aldolase P. stutzeri Hypothetical transmembrane protein Ralstonia solanacearum Glycerophosphoryl diester phosphodiesterase Fusobacterium nucleatum subsp. nucleatum ATCC 25586 Glutamine synthetase, chloroplast precursor Chlamydomonas reinhardtii (eukaryota) Glutamine synthetase, chloroplast precursor C. reinhardtii (eukaryota) UDP-glucose 4-epimerase B. subtilis Glyceraldehyde-3-phosphate dehydrogenase Clostridium perfringens S-adenosylmethionine:tRNA ribosyltransferase-isomerase V. cholerae Malate dehydrogenase Agrobacterium tumefaciens Putative succinyl-CoA synthetase alpha subunit Neisseria meningitidis Z2491

NP 283922 ]

1310 2108

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

163

Table 2. Continued Spot No.

MW (kDa)/pI, theoretical

MW (kDa)/pI, measured

Sequence coverage, %

Protein homologues according to BLAST Protein name Organism

Accession No.

E value

1458

35.5/5.8

35.5/5.7

17

1EVQ A ]

3310 231

1465

35.5/7.7

35.6/8.9

30

NP 404673 ]

1310 2102

1469

33.5/5.1

35.7/5.1

22

NP 246854 ]

2310 275

1512

32.8/5.2

34.3/5.2

16

NP 245081 ]

1310 261

1543

32.2/6.6

33.7/6.5

12

NP 273681 ]

8310 282

1562

32.4/5.7

33.3/5.7

17

NP 248984 ]

6310 289

1586

32.0/7.7

32.2/8.1

25

AAA93461

3310 245

1619

30.1/6.9

31.3/7.7

18

NP 405629 ]

2310 277

1639

30.1/6.9

30.9/9.1

26

NP 405629 ]

2310 277

1653

27.8/5.8

30.3/5.5

19

NP 217921 ]

4310 242

1665

31.3/7.7

30.1/9.1

18

NP 519068 ]

1310 2102

1685

30.3/6.6

29.5/8.4

12

AAK58507

7310 273

1699

27.9/6.2

28.9/5.5

20

NP 396608 ]

3310 211

1709

28.4/6.4

28.9/6.3

28

NP 252347 ]

1310 2102

1715

28.4/6.4

28.8/6.1

28

NP 252347 ]

1310 2102

1724

29.1/5.4

28.3/5.2

19

P22095

1310 288

1749

25.2/5.6

27.4/5.4

31

NP 440208 ]

2310 254

1798

23.7/4.6

25.4/4.7

20

NP 232999 ]

2310 246

1806

26.0/6.1

25.3/6.0

29

NP 233439 ]

2310 284

1815

23.3/6.1

25.1/6.0

28

NP 231650 ]

2310 244

1819

23.4/6.0

25.0/5.9

33

NP 240094 ]

2310 217

1885

21.9/5.4

22.4/5.2

47

NP 456099 ]

1310 273

1887

21.9/5.4

21.9/5.3

47

Chain A, the crystal structure of the thermophilic carboxylesterase Est2 Alicyclobacillus acidocaldarius Acetyl-coenzyme A carboxylase carboxyl transferase subunit alpha Y. pestis FabD P. multocida DdlB P. multocida UTP-glucose-1-phosphate uridylyltransferase N. meningitidis MC58 Probable hydratase P. aeruginosa IMI-1 E. cloacae Septum site-determining protein Y. pestis Septum site-determining protein Y. pestis Hypothetical protein Rv3404c M. tuberculosis H37Rv Probable thymidilate synthase R. solanacearum L-aspartate beta-decarboxylase A. faecalis subsp. faecalis AGR pTi bx104p ] ] A. tumefaciens str. C58 (Cereon) Methionine aminopeptidase P. aeruginosa Methionine aminopeptidase P. aeruginosa Tryptophan synthase alpha chain Vibrio parahaemolyticus Hypothetical protein Synechocystis sp. PCC 6803 Sigma cross-reacting protein 27A V. cholerae Oxidoreductase, short-chain dehydrogenase/reductase family V. cholerae Thymidylate kinase V. cholerae Orotidine 59-phosphate decarboxylase Buchnera sp. APS Superoxide dismutase Salmonella enterica subsp. enterica serovar typhi Superoxide dismutase S. enterica subsp. enterica serovar typhi

NP 456099 ]

1310 273

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

164 Table 2. Continued Spot No.

MW (kDa)/pI, theoretical

MW (kDa)/pI, measured

Sequence coverage, %

Protein homologues according to BLAST Protein name Organism

Accession No.

E value

1943

19.6/4.9

18.8/4.8

17

1JFD A ]

3310 260

1946

19.6/4.9

18.7/4.9

28

CAC92750

3310 260

1968

23.3/5.2

17.8/4.7

20

CAC89968

4310 220

1977

19.7/5.1

17.0/4.9

31

S39907

3310 264

2000

16.3/6.6

15.7/5.3

18

BAB01821

6310 29

2013

15.5/5.9

15.2/5.9

31

NP 297748 ]

2310 254

2019

17.7/4.8

14.8/4.8

41

NP 457678 ]

8310 246

2025

14.0/5.6

14.0/4.8

25

CAB62842

9310 24

2046

16.1/5.6

13.3/5.4

45

F83029

3310 230

2089

11.0/6.7

10.8/6.0

23

T44249

1.2

2119

10.9/5.7

9.9/4.7

13

AAC33847

0.43

2151

12.9/4.6

9.0/4.6

40

Chain A, structure of inorganic pyrophosphatase E. coli Inorganic pyrophosphatase Y. pestis Peptidoglycan-associated lipoprotein Pal Y. pestis Conserved hypothetical protein 10 Rhodobacter capsulatus Alcohol dehydrogenase-like protein Arabidopsis thaliana Nucleoside diphosphate kinase X. fastidiosa 9a5c Transcription elongation factor S. enterica subsp. enterica serovar typhi Hypothetical protein MAL4P2.01 P. falciparum 3D7 50S Ribosomal protein L9 P. aeruginosa Transport protein homolog Arthrobacter sp. strain TE1826 Non-gradient byssal precursor Mytilus edulis 50S Ribosomal protein L7/L12 C. jejuni

NP 281664 ]

3310 225

The mass spectra were recorded on a MALDI mass spectrometer. The amino acid theoretical protein sequences selected in the ORF database by ProteinProspector program were used in the search for proteins with sequence homology of other microbial species listed in the NCBI database using the BLAST program. Theoretical MW/pI values were derived from translated ORFs.

PAGE) has no homology with known proteins. In Table 2 experimental pI and MW values are also compared with theoretical pI and MW values derived from homologues proteins. The major differences between observed and expected molecular mass of proteins concerned proteins with experimental MW below 18 kDa. The major discrepancies in the case of pI values concerned proteins with pI ranging from 8 to 10. The extreme case is membrane protein Tul4 whose theoretical pI value was 9.2 but whose observed value was 4.8. Similar charge and mass variations were described recently for identified proteins of Mycoplasma genitalium [53]. Mostly these discrepancies are associated with post-translational modification of proteins. On the other hand, the possibility of incorrectly identified start and end sites of putative ORFs must be also taken into account.

The reference maps with description of identified proteins will be displayed on a web site, which will allow direct access to 2-DE data (http: / / www.mpiibberlin.mpg.de / 2D-PAGE / microorganisms / index. html).

2.2. Mapping of integral membrane proteins Membrane proteins due to their interfacial position in cells are involved in wide array of host–microbe interactions such as cell adhesion, protein secretion, modulation of cell signaling and induction of endocytosis [54–56]. However, proteomic analysis of membrane proteins was until recently nearly impossible because of technical problems with their extraction and subsequent solubilization [57]. This was especially true for integral membrane proteins that repeatedly traverse membrane bilayer via a

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

helices composed of stretches of hydrophobic amino acids. Now the more effective solubilization of membrane proteins is achieved by introduction of new reagents involving thiourea, tributylphosphine (TBP) and amidosulfobetain surfactants SB3-10 or amidosulfobetain-14 (ASB-14) [58]. Molloy et al. [59] described solubilization and successful 2-DE resolution of some E. coli outer membrane protein using lysis buffer containing 5 M urea, 2 M thiourea, 2% 3-[(3-cholamidopropyl)dimethylamino]-1-propanesulfat (CHAPS), 2% SB3-10 and 2 mM TBP. Further progress made in this field of study coincides with the application of a rapid method for isolation of bacterial outer membrane based on the treatment of isolated bacterial membranes with sodium carbonate at high pH [60]. In this way the cytoplasmic proteins and loosely attached membrane proteins are removed and, conversely, the fraction of integral membrane proteins is enriched. The pellets of integral membrane proteins are then re-solubilized in buffer containing the strong solubilizing additives thiourea, ASB-14 and TBP. The suitability of this approach was verified by the analysis of outer membrane proteome of gram-negative bacteria E. coli and C. crescentus [8,61]. In the case of E. coli, 69% of the integral outer membrane proteins predicted for applied 2-DE separation window were identified on a single gel, while as regards the Caulobacter study, of the 54 proteins uniquely identified, 41 were outer membrane proteins. We have applied the same procedure for detection of F. tularensis LVS outer membrane proteins. The resulting 2-DE map containing more than 300 spots is shown in Fig. 5. Most proteins occur in several charge variants indicating some type of posttranslational modification. It is statistically estimated that |20% of all the identified ORFs in bacteria encode putative integral membrane proteins [58]. In the case of F. tularensis strain Schu 4 ORFeom, 34% of all putative proteins can be tentatively classified as membrane proteins (see below). This assumption roughly corresponds to our number of detected integral membrane proteins [52]. Up to now, only three proteins have been unambiguously identified in this group of proteins. One of them, denoted TUL4, FRATU 17 kDa, has been previously characterized on a molecular level as a 17-kDa major membrane lipoprotein [62]. The other two identified proteins,

165

60-kDa chaperonin and hypothetical 23-kDa protein, were originally thought to reside in cytoplasm. However, concerning chaperonins, several recent studies document their extracytoplasmic location in bacteria [63,64]. Membrane-associated localization of bacterial chaperonins was reported in Salmonella typhimurium [65], Mycobacterium leprae [66], and in several other microbes [67,68]. Membrane forms of chaperonins have different biological functions than their cytoplasmic counterparts. So, while the protection and correct folding of proteins are basic cytoplasmic functions of chaperonins, the membrane chaperonins are implicated in promotion of bacterial adherence and invasiveness [11,69,70]. The Toll-like receptors that are involved in the induction of innate immune response seem to be responsible for interaction with the surface-associated 60-kDa chaperonins [71]. The hypothetical 23-kDa protein is up-regulated during the growth of F. tularensis LVS in macrophages or after in vitro exposure of microbes to hydrogen peroxide [72]. This protein showed no homology with any sequence in databases, and therefore, there has been no indication about its biological function. Nevertheless, since its level is increased under stressful conditions, it can be considered as a universal stress protein. Subcellular fractionation and computer analysis of the amino sequence indicated cytoplasmic location of the hypothetical 23-kDa protein [72]. However, in comparison with published results we identified at least three additional charge variants of the hypothetical 23-kDa protein on our 2-DE protein maps. Hence, like in the case of TUL4 where palmitolaytion of protein was decisive for its membrane location, post-translational modification can also play a role in the hypothetical 23-kDa protein membrane attachment.

2.3. Immunoreactive proteins of F. tularensis LVS microbes F. tularensis infection induces both the cell-mediated and humoral immune response in an infected host. Proteome technology represents a powerful tool for identification of immunologically relevant constituents of F. tularensis. Serological proteome analysis, based on a combination of 2-DE and immunoblotting with human tularemic sera, has the potential to reveal new Francisella specific markers suitable

166

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

Fig. 5. Silver-stained 2-DE reference map of integral membrane proteins of F. tularensis LVS. Isolated membranes were treated with sodium carbonate at high pH. The membrane pellet was solubilized using 5 M urea, 2 M thiourea, 2% CHAPS, 2% SB3-10, 2 mM TBP, 40 mM Tris and 2% carrier ampholytes and proteins were resolved by IEF in the pH range 3–10 followed by SDS–PAGE gradient gel (9–16%). The spot numbers indicate identified proteins.

for diagnostic purposes as well as potentially protective antigens useful for the construction of sub-unit vaccine.

Before the introduction of immunoproteomics, complex antigen preparations such as sonicated whole bacteria [73] or outer membrane preparation

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

containing a variety of proteins [74] were described as immunogens suitable for serodiagnostic tests. Furthermore, Francisella lipopolysaccharide (LPS) also proved to be a diagnostically reliable antigen [75]. The reactivity of immune sera collected from human recipients of the live tularemia vaccine appeared to be directed primarily against this outer membrane component of gram-negative bacteria [76,77]. Similarly, a large portion of the specific antibody response in mice infected with LVS bacteria comprised anti-LPS antibodies [78]. Characteristic LPS ladder-like reactivity can be found on immunoblots treated with both mouse and human immune sera (Fig. 6). As can be seen from this figure, the mouse antibody response exhibiting pronounced reactivity in the low molecular weight region differed clearly from the antibody response elicited in infected human beings (unpublished results). Besides the lipopolysaccharide, other components of F. tularensis outer membrane have been described as immunogens as well. Especially, membrane proteins are potent inducers of specific T cell response in F. tularensis primed humans [79]. One of them, a 17-kDa major membrane lipoprotein denoted TUL4, induced a strong cellular as well as antibody immune response in both vaccinated and naturally infected individuals [80]. Protective immunity against LVS infection has been demonstrated in mice as a consequence of immunization with recombinant S. typhimurium strain expressing TUL4 [81]. A similar protective effect was found after immunization with immunostimulating complexes into which TUL4 had been incorporated [82]. As already mentioned, systemic mapping of the immunoreactive F. tularensis LVS integral membrane proteins, prepared by the carbonate extraction method according to Molloy and coworkers [8], has been started in our laboratory. Using three seropositive tularemic sera, |50 immunoreactive protein spots were detected on 2-D immunoblots [83]. Among them, TUL4 provided strong reaction with all tested sera (Fig. 7). The dnaK and 60-kDa chaperonin are other targets of the humoral antibody response to F. tularensis. The immunoreactivity of these two bacterial stress proteins has been found in sera from naturally infected individuals, as well as in volunteers vaccinated with F. tularensis LVS. On the other hand, weak crossreactivity was also disclosed in some control sera

167

taken from blood donors [84]. This is entirely consistent with our own recently published observations (Table 3) [83]. Furthermore, antibody activity against 10-kDa chaperonin has been demonstrated in this study (Fig. 7). The four charge and mass variants of 10-kDa chaperonin occurred on the 2-DE map of F. tularensis LVS and an additional two 10-kDa chaperonin spots were then identified on the 2-DE map of F. tularensis strain 176 (Table 1, Figs. 1 and 4). The existence of such multiple bacterial variants of 10kDa chaperonin has already been described in a proteome study of M. tuberculosis [85]. According to the obtained results, 10-kDa chaperonin seems to be a promising candidate for Francisella specific marker because of its specificity and immunogenicity (Table 3). Immune reactions with at least one variant of the protein were found in four out of nine individuals undergoing tularemia, whereas no reaction was disclosed in control sera. The list of all immunoreactive proteins identified to date by means of serological proteome analysis is shown in Table 3. In total, 22 immunogenic spots have been successfully identified and 11 of them have provided specific reactions only with sera from tularemia patients. In order to complete the reference map of immunoreactive proteins, several other strongly immunostained spots remain to be identified, especially those on 2-D map of carbonate extracted integral membrane proteins.

2.4. F. tularensis protein profiling under conditions mimicking a hostile intracellular environment When bacteria enter the host organism they undergo phenotypic modulation controlled by the coordinated gene expression in order to survive harsh conditions and hostile environmental conditions [86]. Moreover, intracellular parasites, which invade and multiply within cells, specialized to their ingestion and killing, must resist intracellular defensive mechanisms of both oxidative and non-oxidative origin. The oxygen dependent mechanisms rely on the generation of reactive oxygen metabolites, such as superoxide anion, hydroxyl radicals, hypochlorite ions, hydrogen peroxide, and singlet oxygen inside the phagosome. The acidification of intraphagosomal

168

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

Fig. 6. 2-D immunoblotting pattern of specific antibody response to F. tularensis infection. The immune sera were collected from C3H / HeN (Lps n) mice (A), and from their congenic C3H / HeJ (Lps d ) counterparts (B), 28 days after F. tularensis LVS challenge. Mice were inoculated s.c. with a dose of 100 microbes. Non-immune control sera originated from mice treated only with saline. For comparison, the representative human immunoreactivity pattern is shown (C). Human tularemic serum was obtained from a patient with natural infection. As a control, serum from a healthy blood donor without a history of tularemia was used. The whole cell bacterial lysate of F. tularensis LVS was separated on pH 3–10 IPG strips, followed by 9–16% SDS–PAGE. All the sera were diluted 1:100. As a secondary antibody anti-IgG, IgA, and IgM peroxidase conjugate was used. Reactions were visualized with enhanced chemiluminiscence based kit.

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

169

Fig. 7. An example of a 2-D immunoblotting pattern of antigens recognized by human tularemic serum. Whole cell bacterial lysate F. tularensis LVS (A) and integral membrane proteins, isolated by the carbonate extraction method according to Molloy and coworkers [8] (B), were resolved on pH 3–10 IPG strips with subsequent 9–16% SDS–PAGE [84]. Both blotting membranes were treated with the same tularemic serum. As a secondary antibody anti-IgG, IgA, and IgM peroxidase conjugate was used. Reactions were visualized with enhanced chemiluminiscence based kit.

space, release of hydrolytic enzymes and small defensine molecules, and deprivation of nutrients are then examples of cellular oxygen-independent defensive potential. Actually each stress factor induces a specific set of proteins. Therefore, the proteomic analysis of changes in bacterial gene expression under the influence of real intracellular conditions or conditions mimicking an adverse intracellular milieu offers the unique possibility of identifying molecules responsible for virulence and pathogenicity of intracellular bacterial pathogens [87]. To determine the pattern of bacterial gene expression in terms of defined responses to conditions relevant to the intracellular environment, several in vitro systems were exploited. The cultivation of bacteria under selected stress conditions, e.g. low or high concentrations of iron; extreme temperatures,

acidic pH, or oxidative stress, were used for the mapping of molecules significant for microbial pathogenicity. The most sophisticated system for this type of study seems to be the radiolabeling of bacterial proteins after in vitro ingestion of microbes by phagocytic cells. The utilization of this complex in vitro system reflects the facts documenting that the bacterial response to intracellular growth is not simply the sum of responses to individual stress conditions mentioned above [21]. Currently there is little data concerning the modulation of F. tularensis gene expression in the course of microbe–host cell interaction. Analysis of protein synthesis of F. tularensis LVS growing intracellularly in the macrophage-like murine J774 macrophages demonstrated induction of only a few proteins [72]. One-dimensional electrophoresis of bacteria pulse-

170

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

Table 3 The list of identified F. tularensis LVS immunoreactive proteins Protein Protein name Organism

Chaperone protein dnaK F. tularensis 60-kDa Chaperonin F. tularensis Elongation factor TU S. typhimurium Glycine-cleavage system protein T1 P. aeruginosa Oxidoreductase V. cholerae Hypothetical protein P. falciparum Hypothetical 23-kDa protein F. tularensis Biotin carboxyl carrier protein P. aeruginosa 50S Ribosomal protein L9 P. aeruginosa Probable bacterioferritin P. aeruginosa 10-kDa Chaperonin F. tularensis 10-kDa Chaperonin F. tularensis 50S Ribosomal protein C. jejuni 10-kDa Chaperonin F. tularensis 3-Dehydroquinase Buchnera sp. 10-kDa Chaperonin F. tularensis

Histone-like protein HU form B P. aeruginosa

60-kDa Chaperonin F. tularensis 17-kDa Major membrane protein (precursor) F. tularensis

2-D immunoblotting reactivity Accession No.

P48205

Spot No.

MW (kDa) / pI, measured

Whole-cell antigen F. tularensis LVS IEF pH 3–10 9–16% SDS–PAGE 120 69.2 / 4.9

Tularemia

Borreliosis

Blood donors

111

1/2

1

P94798

214

55.6 / 4.9

111

1

–

P21694

340

47.8 / 4.6

111

111

–

G82994

488

40.3 / 5.9

111

–

–

G82383

753

25.7 / 6.3

1

–

–

CAB38995

755

25.6 / 5.3

1

–

–

CAA70085

790

23.3 / 5.4

1/2

1

1/2

P37799

889

13.7 / 4.9

111

1/2

–

F83029

893

13.3 / 5.4

–

1

1

B83036

918

10.3 / 5.4

111

1/2

–

P94797

920

10.3 / 5.2

11

–

–

P94797

931

9.7 / 4.9

111

–

–

H81392

933

9.6 / 4.6

111

11

111

P94797

939

9.3 / 5.3

111

–

–

P57479

949

9.4 / 6.0

11

–

–

P94797

962

8.6 / 5.3

11

–

–

Whole-cell antigen F. tularensis LVS IEF pH 6–11 9–16% SDS–PAGE AF345628 478 14.5 / 9.1

11

1/2

–

Integral membrane proteins F. tularensis LVS IEF pH 3–10 9–16% SDS–PAGE P94798 47 57.4 / 5.0

111

–

–

P18149

111

1

–

275

13.3 / 4.8

Human tularemic sera originating from nine seropositive patients were used to screen for the immunoreactive antigens of F. tularensis by the means of 2-D immunoblotting [83]. As controls, sera from two blood donors and three individuals with Lyme disease were used. Control sera were anti-F. tularensis negative according to microagglutination, but displayed some immunoreactivity with F. tularensis antigens on 1-D immunoblots. Maximum intensity of immunostaining found on 2-D immunoblots is shown in the table. The staining intensity was evaluated visually and classified from 2 to 111. Either MALDI–TOF-MS (whole cell lysate), or LC–MS–MS (integral membrane proteins) was used for identification. Both the F. tularensis proteins included in the NCBI or SWISSPROT protein databases, as well as sequence homologues originating from other bacteria, are presented.

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

labeled at 24 h of intracellular growth revealed the increased synthesis of four proteins, of which only one protein, hypothetical 23-kDa protein, was shown to be also induced using 2-DE gels. Surprisingly, with the exception of dnaK, the 60- and 10-kDa chaperonins were only marginally or not at all induced. Upregulation of the hypothetical 23-kDa protein was also found after exposure of F. tularensis LVS to hydrogen peroxide. In contrast, the temperature shift to 42 8C that led to the increased expression of at least 15 proteins, involving dnaK, 60- and 10-kDa chaperonins, did not influence the hypothetical 23-kDa protein level significantly [72].

3. Comparative analysis of F. tularensis type A and B strains A unique set of phenotypic characteristics, including morphology and biochemical markers, made isolates of the Francisella genus readily distinguishable. However, within the genus and particularly within species F. tularensis, the strain discrimination is not performed conveniently. Previous genetic studies have demonstrated that the four subspecies of F. tularensis, despite showing marked variations in their virulence for mammals and originating from different regions throughout the northern hemisphere, display a very close phylogenetic relationship. Due to the contagiousness of Francisella isolates only limited work has been performed to develop typing methods based on cultivation. Recent development in genetic analysis made novel strategies based on PCR discrimination available. Johansson and his group [88] tested three PCR methods which all generated similar subspecies specific patterns, however, the discriminatory indices varied within the range, which is not sufficient to distinguish individual strains. They proposed the combination of one of these methods and PCR employing subspecies-specific primers as a suitable, rapid and relatively simple strategy for discrimination of Francisella species and subspecies. Identification of subspecies specific markers using whole-genome DNA micro array and subsequent clustering reveals similar data regarding subspecies distribution [34]. Both results suggest that Japanese strains sorted to subsp. holarctica represent a separate subspecies and

171

that type strain ATCC6223 differ from other strains, which authors explain by genetic changes during storage and passages in small animals. The DNA micro array study shows an almost identical pattern of subsp. mediaasiatica and subsp. tularensis which contradicts current strain designation. PCR study does not support this result. A thorough comparative proteome analysis of different strains is under way. The 19 strains included so far in the study belong to three subspecies of F. tularensis—tularensis, holarctica and mediaasiatica (Table 4). The 2-DE map of whole cell lysate of each strain has been performed on a wide gradient of pH 3–10 (IPG strips) and spot patterns of all gels have been compared using Melanie III software. With preliminary comparison, an automatic correspondence analysis (Melanie III) has been performed after landmark identification and gel alignment. Correspondence analysis allows representation of series of gels, each having a large number of spots, in a factorial space of reduced dimension. Gels appear as points on the planar graphic plot and the distance between points corresponds to their similarity [89]. This analysis applied to 19 images of gels revealed a distinct pattern of distribution. Whilst all holarctica strains were grouped together, tularensis strains scatter more dramatically, however, they were clearly distinguished from holarctica strains (unpublished results). The most diverse locations exhibit tularensis strains harboring resistance to antibiotics (FAM SR which is streptomycin resistant, and TTCR 6-4-1 which is tetracycline resistant) and ATCC tularensis type strain ATCC6223. The separation of ATCC variant is in accordance with the results from both genetic studies mentioned above and indicates the loss of some characteristics typical for wild tularensis isolates. The distribution of antibiotic resistant strains suggests large changes in protein profiles associated with the induction of antibiotic resistance and presents an interesting task for further analyses. The separate grouping of two mediaasiatica strains with respect to both tularensis and holarctica subsp. supports their distinct subspecies annotation, although this result needs further evaluation. For the purpose of comparative proteomic study of subsp. tularensis and holarctica, the strain selection reflected correspondence analysis distribution (the

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

172

Table 4 F. tularensis strains included in comparative analysis FSC No.a

Subspecies

Origin

Other names

Selected b

13 14 41 43 122 199 230 237 17 35 69 74 124 155 200 247 257 147 148

tularensis tularensis tularensis tularensis tularensis tularensis tularensis tularensis holarctica holarctica holarctica holarctica holarctica holarctica holarctica holarctica holarctica mediaasiatica mediaasiatica

(from Eigelsbach) Strepto resistant FAM-variant Tick, British Columbia, Canada, 1935 Human ulcer, 1941, Ohio Ttc resistant Schu-strain Mite, Slovakia, 1988 Human lymph node, 1920, Utah Human ulcer, 1941, Ohio Human lymph node, 1926, Japan Beaver 1976, Hamilton, Montana T7 passaged in animals, more virulent Hare, Sweden, 1974 Water, 1990, Odessa region, Ukraine Live Vaccine Strain, LVS, Russia Human, 1998, Ljusdal, Sweden Human, 1993, Vosges, France Tick, Moscow area, 1949, Russia Miday gerbil, 1965, Kazakhstan Ticks, 1982, Central Asia

FAM standard FAM SR Vavenby Schu TTCR 6-4-1 SE-221-38 ACC 6223 Schu-S4 Jap-S2 B423 A SVA T7K SVA T7 14588 ATCC 29684 Rem nr 3001 SVA T20 503 / 840 543 240

Yes No Yes Yes No Yes No Yes No Yes No Yes No No Yes Yes Yes – –

The 2-DE map of whole cell lysate on wide range IPG strip (pH 3–10) of each strain has been prepared and correspondence analysis performed. Subsequently, strains for holarctica versus tularensis comparison have been selected based on 2-DE map similarity within subspecies designation. a Francisella Strain Collection number at SDRA. b Strains selected for holarctica versus tularensis comparison.

most related strains were selected). Five strains, each represented by one gel, of subsp. tularensis and five strains, each represented by one gel, of subsp. holarctica were included (Table 4). Out of the total of 69 protein spot differences, 30 were unique to tularensis strains and 39 to holarctica strains. The identification of the specific protein by peptide mass fingerprinting and other mass spectrometry based methods is ongoing and, currently, seven differentially expressed proteins have been identified (Table 5). Two proteins specific for subsp. tularensis, the most basic variant of hypothetical 23-kDa protein and acid phosphatase (Acp), have been identified previously [90]. As described by Golovliov et al. [72], hypothetical 23-kDa protein was the most upregulated protein during intracellular growth of F. tularensis vaccine strain LVS in macrophages, pinpointing this protein as one of the possible virulence factors. Acid phosphatase specificity corresponds with the common idea that this enzyme belongs to a group of virulence factors of intracellular pathogens. Never-

theless, some unique features in the structure and substrate specificity of tularemic Acp as compared to other burst-inhibiting Acps makes it difficult to fit this protein into currently recognized classes of acid phosphatases [91] and thus makes explanation of the mechanism leading to increased pathogenicity unclear. Out of 39 proteins specific for holarctica strains, five spots have been identified so far. Among them the expression of bacterioferritin (Fig. 8) exhibited the most striking difference [23]. Recently published work aimed at comparative protein profiling of vaccine and virulent strains of Brucella melitensis [92] described a similar finding with characteristic expression of bacterioferritin in vaccine strain. In this study, not only bacterioferritin but also two other proteins involved in iron transport and utilization have been upregulated in vaccine strain. Deprivation of iron is an important factor of host defense against pathogenic bacteria and the battle for iron is one of the most important microbe–host interactions [93].

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

173

Table 5 The list of identified proteins with unique expression in holarctica or tularensis subspecies Protein name Organism Specific for holarctica Probable bacterioferritin P. aeruginosa Probable sigma (54) modulation protein P. putida Putative succinyl-CoA synthetase alpha subunit Neisseria meningitidis Z2491 Nucleoside diphosphate kinase Xylella fastidiosa 9a5c Cationic 19-kDa outer membrane protein precursor Yersinia Specific for tularensis Acid phosphatase F. tularensis Hypothetical 23-kDa protein F. tularensis

Accession No.

MW (kDa) / pI, theoretical

MW (kDa) / pI, measured

Sequence coverage, %

B83036

18.5 / 5.3

10.3 / 5.4

40

P15592

11.2 / 5.9

7.8 / 6.3

46

NP-283922

30.1 / 6.1

36.1 / 5.9

21

NP-297748

15.5 / 5.9

15.2 / 5.9

31

P315119

18.8 / 7.7

14.4 / 6.4

34

AAB06624

57.6 / 5.9

59.0 / 6.0

23

CAA70085

22.1 / 5.7

23.0 / 6.0

30

Specific spots for each subspecies have been selected based on 2-DE map comparison of selected strains. Out of 69 specific spots, seven proteins, mentioned in the table, have been identified so far by comparing spot position to position of proteins already identified by peptide mass fingerprinting in 2-DE database of F. tularensis LVS (subsp. holarctica) and F. tularensis 176 (subsp. tularensis).

4. Construction of virtual and real proteome of F. tularensis—from genome to proteome In order to verify and extend our knowledge of newly identified F. tularensis proteins, we have

analyzed translated ORF of each identified protein by algorithms providing information about functional and topological protein classification. The computing working on the basis of sequence similarity is able to discriminate between soluble and membrane proteins

Fig. 8. Enlargement of a section of the 2-DE map showing different expressions of bacterioferritin (arrow) between subsp. tularensis (FSC 199) and holarctica (FCS 74). Silver stained 2-DE gels, pH 3–10, 9–16% SDS–PAGE.

174

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

(SOSUI), to predict intracellular topology (PSORT) of proteins and to determine the tentative function of proteins (COG algorithm). The data obtained from this analysis are being used for the formation of a so-called ‘‘virtual proteome’’ of F. tularensis microbe. A software system, SOSUI, was developed for discriminating between soluble and membrane proteins together with the prediction of transmembrane helices [94]. The performance of the system was 99% accurate for discrimination between two types of proteins and 96% for prediction of transmembrane helix [95]. Preliminary analysis of F. tularensis ORFeom by the SOSUI system predicted that 34% of all putative proteins could be located in bacterial membrane. However, the data should be taken with caution because prediction of membrane location with reasonable confidence necessitates the evaluation of additional criteria such as sequence similarities to other membrane proteins and the existence of signal sequence for translocation of protein to outer membrane. For this purpose the localization of proteins was also verified by the PSORT algorithm which was developed to predict protein topology inside gram-positive and gram-negative bacteria, yeast, animal and plant cells. The predicted candidate localization-sites for gram-negative bacterium are bacterial outer membrane, bacterial periplasmic space, bacterial inner membrane and bacterial cytoplasm. PSORT first predicts the presence of signal sequences by the method of McGeoch [96] modified by Nakai and Kanehisa [97]. Further, PSORT applies von Heijne’s method [98] of signal sequence recognition. The next parameters that PSORT calculates are transmembrane segments and lipoproteins according the methods published earlier [99,100]. The functional classification of tentative proteins was obtained by applying the Clusters of Orthologous Groups of proteins (COGs) algorithm. The COG database has been developed as a phylogenetic classification of proteins from complete genomes [101]. The COGs have been classified into 17 broad functional categories, including a class for which a general function prediction, usually that of biochemical activity, was feasible and a class of uncharacterized COGs. In a strict sense, the COGNITOR program which accompanies the COG database and assigns new proteins [102] was used for prediction of identified F. tularensis proteins function. The ana-

lyzed data covering both identified proteins with known function and hypothetical proteins are shown in Table 6. As can be seen from this table, proteins identified by mass spectrometry as enzymes participating in the metabolism of amino acids, carbohydrates, lipids, nucleotides or energy production, have been sorted into the COG category that corresponds with their assigned function. This further confirms the reliability of the identification approach. The same is true for chaperones, transcription and translation factors. Topological prediction was, for all the above-mentioned functional groups, besides chain A, structure of inorganic pyrophosphatase, determined as bacterial cytoplasm protein. The function of remaining identified proteins listed in Table 6 was unknown. Two proteins, peptidoglycan-associated lipoprotein Pal and IMI-1, were classified by COG in the COG-M category, i.e. cell envelope biogenesis, outer membrane which is also in agreement with their predicted topological location (bacterial inner membrane proteins containing signal peptides). The classification of hypothetical transmembrane protein was COG-P-inorganic ion transport and metabolism. Both algorithms SOSUI and PSORT predicted its membrane position. For five protein classifications COG-S-function was unknown or no related COG were found. For the latter, the group includes two membrane proteins not described before and the already mentioned TUL4 and hypothetical 23-kDa proteins. The topography prediction of TUL4 documents the necessity to combine SOSUI and PSORT algorithms. As was stated, TUL4 is integral membrane protein anchored in membrane via palmitic acid. Because of the lack of transmembrane domains, SOSUI recognized it as soluble protein, however, by the PSORT algorithm it was designated as an outer membrane protein.

5. Conclusions The current proteomic approach still suffers from several technical limitations that do not allow comprehensive analysis of the expression of many important classes of proteins to be performed. However, despite this, even initial bacterial proteome analyses are already able to yield important data contributing to a more detailed understanding of the pathogenesis of infectious diseases. This further

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

175

Table 6 Prediction of membrane association using programs SOSUI and PSORT, and functional assignment of genes based on placing ORF in COG category Protein name (homologue)

NAD-specific glutamate dehydrogenase Glutamine synthetase, chloroplast precursor Phosphoglycerate kinase Fructose-bisphosphate aldolase Peptidoglycan-associated lipoprotein Pal IMI-1 Citrate synthase Chain A, structure of inorganic pyrophosphatase Sigma cross-reacting protein 27A Macrophage growth locus B Hypothetical transmembrane protein Probable bacterioferritin FabD Transport protein homolog Non-gradient byssal precursor Hypothetical 23-kDa protein TUL4, FRATU 17 kDa Thymidylate kinase Probable thymidylate synthase Chaperone protein dnaK 60-kDa Chaperonin 10-kDa Chaperonin Transcription elongation factor 50S Ribosomal protein L9

Structure–function relationship (COG-code)

Amino acid transport and metabolism (COG-E) Amino acid transport and metabolism (COG-E) Carbohydrate transport and metabolism (COG-G) Carbohydrate transport and metabolism (COG-G) Cell envelope biogenesis, outer membrane (COG-M) Cell envelope biogenesis, outer membrane (COG-M) Energy production and conversion (COG-C) Energy production and conversion (COG-C) Function unknown (COG-S) General function prediction only (COG-R) Inorganic ion transport and metabolism (COG-P) Inorganic ion transport and metabolism (COG-P) Lipid metabolism (COG-I) No related COG No related COG No related COG No related COG Nucleotide transport and metabolism (COG-F) Nucleotide transport and metabolism (COG-F) Posttranslation modification, protein turnover, chaperons (COG-O) Posttranslation modification, protein turnover, chaperons (COG-O) Posttranslation modification, protein turnover, chaperons (COG-O) Transcription (COG-K) Translation, ribosomal structure and biogenesis (COG-J)

encourages rapid progress in improvements of technologies for protein analysis that should in combination with other complementary techniques, such as molecular biology, enable complex microbial systems to be studied in their entirety.

Acknowledgements The authors wish to acknowledge the excellent technical support provided by Jana Michalickova and Alena Firychova. This work was supported by Grant LN00A033 from the Ministry of Education, Youth and Sport, Czech Republic.

References ¨ [1] A. Gorg, Nature 349 (1991) 545. [2] M.P. Washburn, D. Wolters, J.R. Yates, Nat. Biotechnol. 19 (2001) 242.

Topography prediction

Signal

SOSUI

PSORT

peptide

Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Membrane protein Soluble protein Soluble protein Membrane protein Membrane protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein Soluble protein

Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial inner membrane protein Bacterial inner membrane protein Bacterial cytoplasm protein Bacterial inner membrane protein Bacterial inner membrane protein Bacterial cytoplasm protein Bacterial inner membrane protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial inner membrane protein Bacterial inner membrane protein Bacterial cytoplasm protein Bacterial outer membrane protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial cytoplasm protein Bacterial cytoplasm protein

No No No No Yes Yes No No No No Yes Yes No Yes Yes No Yes No No No No No No No

[3] S.J. Cordwell, A.S. Nouwens, B.J. Walsh, Proteomics 1 (2001) 461. [4] L. Tonella, C. Hoogland, P.A. Binz, R.D. Appel, D.F. Hochstrasser, J.C. Sanchez, Proteomics 1 (2001) 409. [5] H. Langen, B. Takacs, S. Evers, P. Berndt, H.W. Lahm, B. Wipf, C. Gray, M. Fountoulakis, Electrophoresis 21 (2000) 411. [6] H.J. Mollenkopf, P.R. Jungblut, B. Raupach, J. Mattow, S. Lamer, U. Zimny-Arndt, U.E. Schaible, S.H.E. Kaufmann, Electrophoresis 20 (1999) 2172. [7] D. Bumann, T.F. Meyer, P.R. Jungblut, Proteomics 1 (2001) 473. [8] M.P. Molloy, B.R. Herbert, M.B. Slade, T. Rabilloud, A.S. Neuwens, K.L. Williams, A.A. Gooley, Eur. J. Biochem. 267 (2000) 2871. [9] V. Santini, S. Kieffer, D. Desclaux, F. Masson, T. Rabilloud, Electrophoresis 21 (2000) 3329. [10] P.A. Cullen, S.J. Cordwell, D.M. Bulach, D.A. Haake, B. Adler, Infect. Immunol. 70 (2002) 2311. [11] N.D. Phadke, M. Molloy, S.A. Steinhoff, P.J. Ulintz, P.C. Andrews, J.R. Maddock, Proteomics 1 (2001) 705. ¨ [12] S. Backert, E.C. Muller, P.R. Jungblut, T.F. Meyer, Proteomics 1 (2001) 608. [13] K. Weldingh, I. Rosenkrands, S. Jacobsen, P.B. Rassmussen, M.J. Elhay, P. Andersen, Infect. Immunol. 66 (1998) 3492. [14] I. Hirose, K. Samo, I. Shioda, M. Kumano, K. Nakamura, K. Yamane, Microbiology 146 (2000) 65.

176

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek

[15] P.R. Jungblut, H.U.E. Schaible, J. Mollenkopf, U. ZimnyArndt, B. Raupach, J. Mattow, P. Halada, S. Lamer, K. Hagens, S.H.E. Kaufmann, Mol. Microbiol. 33 (1999) 1103. [16] S.L. Hanna, N.E. Sherman, M.T. Kinter, J.B. Goldberg, Microbiology 146 (2000) 2495. [17] D.K. Wong, B.Y. Lee, M.A. Horwitz, B.W. Gibson, Infect. Immunol. 67 (1999) 327. [18] A.P. Teixeira-Gomez, A. Cloeckaert, M.S. Zygmunt, Infect. Immunol. 68 (2000) 2954. [19] L. Phan-Thanh, F. Mahouin, Electrophoresis 20 (1999) 2214. [20] J.A. Streit, J.E. Donelson, M.W. Agey, M.E. Wilson, Infect. Immunol. 64 (1996) 1810. [21] M. Rafie-Kolpin, R.C. Essenberg, J.H. Wyckoff, Infect. Immunol. 64 (1996) 5274. [22] R.C. Fernandez, S.M. Logan, S.H.S. Lee, P.S. Hoffman, Infect. Immunol. 64 (1996) 1968. [23] G. Haas, G. Karaeli, K. Ebermayer, W.G. Metzger, S. Lamer, U. Zimny-Arndt, S. Diescher, U.B. Goebel, K. Vogt, A.B. Roznowski, B.J. Wiedenmann, T.F. Meyer, T. Aebischer, P.R. Jungblut, Proteomics 2 (2002) 313. [24] B.A. Covert, J.S. Spencer, I.M. Orme, J.T. Belisle, Proteomics 1 (2001) 574. [25] G.W. McCoy, Public Health Bull. 43 (1911) 53. [26] G.W. McCoy, C.W. Chapin, J. Infect. Dis. 10 (1912) 61. [27] E. Francis, J. Am. Med. Assoc. 84 (1925) 1243. [28] H.T. Eigelsbach, V.G. McGann, in: M.P. Starr, H. Stolp, H.G. ¨ Truper, A. Balows, H.G. Schlegel (Eds.), The Prokaryotes, Vol. I, Springer, Berlin, 1981, p. 1086. [29] B. Lukas, J. Libich, Cz. Epidemiol. Mikrobiol. Immunol. 11 (1967) 290. [30] N.G. Olsufjev, O.S. Emelyanova, T.N. Dunayeva, J. Hyg. Epidemiol. Microbiol. Immunol. (Prague) 3 (1959) 138. [31] W.L. Jellison, C.R. Owen, J.F. Bell, G.M. Kohls, Wildlife Dis. 17 (1961) 1. [32] N.G. Olsufjev, I.S. Meshcheryakova, J. Hyg. Epidemiol. Microbiol. Immunol. (Prague) 26 (1982) 291. ¨ A. Sjostedt, ¨ [33] G. Sandstrom, M. Forsman, N.V. Pavlovich, B.N. Mishankin, J. Clin. Microbiol. 30 (1992) 172. ¨ [34] M. Broekhuijsen, P. Larsson, A. Johansson, A. Sjostedt, R.G. Prior, R.W. Titball, M. Forsman, in: Proceedings of the VIIth International BC-Protection Symposium, Stockholm, 2001. [35] N.A. Gajskij, Zh. Mikrobiol. Epidemiol. Immunobiol. 15 (1944) 14. [36] W.D. Tigertt, Bacteriol. Rev. 26 (1962) 354. [37] H.T. Eigelsbach, C.M. Downs, J. Immunol. 87 (1961) 415. ¨ J. Chem. Technol. Biotechnol. 59 (1994) 315. [38] G. Sandstrom, [39] A. Macela, Thesis, Purkyneˇ Military Medical Academy, ´ ´ 1980. Hradec Kralove, [40] J. Libich, Tularemia, Avicenum, Prague, 1981. [41] D.S. Burke, J. Infect. Dis. 135 (1977) 55. [42] A. Macela, H. Kovarova, J. Kopecky, P. Propper, in: M. Ryc, J. Franek (Eds.), Bacteria and the Host, Avicenum, Prague, 1986, p. 287. [43] L.S.D. Anthony, P.A.L. Kongshavn, Microb. Pathog. 2 (1987) 3. [44] A.H. Fortier, D.A. Leiby, R.B. Narayan, E. Asafoadjei, R.M. Crawford, C.A. Nacy, M.S. Meltzer, Infect. Immunol. 63 (1995) 1478.

[45] L.S.D. Anthony, R.D. Burke, F.E. Nano, Infect. Immunol. 59 (1991) 3291. ¨ [46] X.-H. Lai, I. Golovliov, A. Sjostedt, Infect. Immunol. 69 (2001) 4691. [47] L. Hernychova, H. Kovarova, A. Macela, M. Kroca, Z. Krocova, J. Stulik, Immunol. Lett. 57 (1997) 75. [48] R. Karttunen, H.M. Surcel, G. Andersson, H.P. Ekre, E. Herva, J. Clin. Microbiol. 29 (1991) 753. [49] T. Rabilloud, C. Adessi, A. Giraudel, J. Lunardi, Electrophoresis 18 (1997) 307. [50] J.-C. Sanchez, V. Rouge, M. Pisteur, F. Ravier, L. Tonella, M. Moosmayer, M.R. Wilkins, D.F. Hochstrasser, Electrophoresis 18 (1997) 324. ¨ [51] H. Schagger, G. Jagow, Anal. Biochem. 166 (1987) 368. [52] R.G. Prior, L. Klasson, P. Larsson, K. Williams, L. Lindler, ¨ A. Sjostedt, T. Swensson, I. Tamas, K.W. Wren, P.C. Oyston, S.G. Andersson, K.W. Titball, J. Appl. Microbiol. 4 (2001) 614. [53] V.C. Wasinger, J.D. Pollack, I. Humphrey-Smith, Eur. J. Biochem. 267 (2000) 1571. [54] D.M. Underhill, A. Ozinsky, Annu. Rev. Immunol. 20 (2002) 825. [55] M. Rescigno, G. Rotta, B. Valzasina, P. Ricciardi-Castagnoli, Immunobiology 204 (2001) 572. [56] D.M. Underhill, A. Ozinski, Curr. Opin. Immunol. 14 (2002) 103. [57] V. Santoni, S. Kieffer, D. Desclaux, F. Masson, T. Rabilloud, Electrophoresis 21 (2000) 3329. [58] V. Santoni, M. Molloy, T. Rabilloud, Electrophoresis 21 (2000) 1054. [59] M. Molloy, B.R. Herbert, B.J. Walsh, M.I. Tyler, M. Traini, J.-C. Sanchez, D.F. Hochstrasser, K.L. Williams, A.A. Gooley, Electrophoresis 19 (1998) 1960. [60] Y. Fujiki, A.L. Hubbard, S. Fowler, P.B. Lazarow, J. Cell Biol. 93 (1982) 97. ¨ ¨ ¨ Infect. Immunol. 59 [61] A. Sjostedt, A. Tarnvik, G. Sandstrom, (1991) 3163. ¨ ¨ I. Horvath, L. Vigh, Plant Physiol. [62] E. Kovacs, Z. Torok, Biochem. 32 (1994) 285. [63] A. Scorpio, A. Johnson, A. Laquerre, D.R. Nelson, J. Bacteriol. 176 (1994) 6449. [64] M. Ensgraber, M. Loos, Infect. Immunol. 60 (1992) 3072. [65] T.P. Gillis, R.A. Miller, D.B. Young, S.R. Khanolkar, T.M. Buchanan, Infect. Immunol. 49 (1985) 371. [66] H. Yamaguchi, T. Osaki, H. Taguchi, T. Hanawa, T. Yamamoto, S. Kamiya, J. Gastroenterol. 31 (1996) 12. [67] B.B. Plikaytis, G.M. Carlone, C.-P. Pau, H.W. Wilkinson, J. Clin. Microbiol. 25 (1987) 2080. [68] H. Yamaguchi, T. Osaki, N. Kurihara, H. Taguchi, T. Hanawa, T. Yamamoto, S. Kamiya, J. Med. Microbiol. 46 (1997) 825. [69] C. Hennequin, F. Porcheray, A. Dupriet-Waligora, A. Colignon, M. Barc, P. Bourlioux, T. Karjalainen, Microbiology 147 (2001) 87. [70] R.A. Garduno, E. Garduno, P.S. Hoffman, Infect. Immunol. 66 (1998) 4602. [71] S. Sasu, D. LaVerda, N. Qureshi, D.T. Golenbock, D. Beasley, Circ. Res. 89 (2001) 244.

´ et al. / J. Chromatogr. B 787 (2003) 149–177 M. Hubalek ¨ ¨ [72] I. Golovliov, M. Eriksson, G. Sandstrom, A. Tarnvik, A. ¨ Sjostedt, Infect. Immunol. 65 (1997) 2183. [73] M.K. Viljanen, T. Nurmi, A. Salminen, J. Infect. Dis. 148 (1983) 15. [74] L. Bevanger, J.A. Maeland, A.I. Naess, J. Clin. Microbiol. 26 (1988) 433. [75] H.E. Carlsson, A.A. Lindberg, G. Lindberg, B. Hederstedt, K.A. Karlsson, B.O. Agell, J. Clin. Microbiol. 10 (1979) 615. ¨ [76] D.M. Waag, A. Galloway, G. Sandstrom, C.R. Bolt, M.J. England, G.O. Nelson, J.C. Williams, J. Clin. Microbiol. 30 (1992) 2256. [77] J.J. Drabick, R.B. Narayanan, J.C. Williams, J.W. Leduc, C. Nacy, Am. J. Med. Sci. 308 (1994) 83. [78] V.C. Dreisbach, S. Cowley, K.L. Elkins, Infect. Immunol. 68 (2000) 1988. ¨ A. Tarnvik, ¨ [79] G. Sandstrom, H. Wolf-Watz, J. Clin. Microbiol. 25 (1987) 641. ¨ ¨ ¨ [80] A. Sjostedt, G. Sandstrom, A. Tarnvik, B. Jaurin, J. Immunol. 145 (1990) 311. ¨ ¨ A. Tarnvik, ¨ [81] A. Sjostedt, G. Sandstrom, Infect. Immunol. 60 (1992) 2855. ˚ ¨ [82] I. Golovliov, M. Ericsson, L. Akerblom, G. Sandstrom, A. ¨ ¨ Tarnvik, A. Sjostedt, Vaccine 139 (1995) 261. [83] J. Havlasova, L. Hernychova, P. Halada, V. Pellantova, J. Krejsek, J. Stulik, A. Macela, P.R. Jungblut, P. Larsson, M. Forsman, Proteomics 2 (2002) 857. ¨ ¨ [84] M. Ericsson, I. Golovliov, G. Sandstrom, A. Tarnvik, A. ¨ Sjostedt, Infect. Immunol. 65 (1997) 1824. [85] J. Mattow, P.R. Jungblut, E.-C. Muller, S.H.E. Kaufmann, Proteomics 1 (2001) 494. [86] M. Rhen, S. Eriksson, S. Petersson, Curr. Opin. Microbiol. 3 (2000) 60.

177

[87] P. Cash, Electrophoresis 21 (2000) 1187. [88] A. Johansson, A. Ibrahim, I. Goransson, U. Eriksson, D. ¨ Gurycova, J.E. Clarridge III, A. Sjostedt, J. Clin. Microbiol. 38 (2000) 4180. [89] T. Pun, D.F. Hochstrasser, R.D. Appel, M. Funk, V. VillarsAugsburger, C. Pellegrini, Appl. Theor. Electrophoresis 1 (1988) 3. [90] L. Hernychova, J. Stulik, P. Halada, A. Macela, M. Kroca, T. Johansson, M. Malina, Proteomics 1 (2001) 508. [91] T.J. Reilly, G.S. Baron, F.E. Nano, M.S. Kuhlenschmidt, J. Biol. Chem. 271 (1996) 10973. [92] M. Eschenbrenner, M.A. Wagner, T.A. Horn, J.A. Kraycer, C.V. Mujer, S. Hagius, P. Elzer, V.G. DelVecchio, J. Bacteriol. 184 (2002) 4962. [93] C. Ratledge, L.G. Dover, Annu. Rev. Microbiol. 54 (2000) 881. [94] T. Hirokawa, S. Boon-Chieng, S. Mitaku, Bioinformatics 14 (1998) 378. [95] S. Mitaku, M. Ono, T. Hirokawa, S. Boon-Chieng, M. Sonoyama, Biophys. Chem. 82 (1999) 165. [96] D.J. McGeoch, Virus Res. 3 (1985) 271. [97] K. Nakai, M. Kanehisa, Proteins 11 (1991) 95. [98] G. von Heijne, Nucleic Acids Res. 14 (1986) 4683. [99] P. Klein, M. Kanehisa, C. DeLisi, Biochim. Biophys. Acta 815 (1985) 468. [100] K. Yamaguchi, F. Yu, M. Inoue, Cell 53 (1988) 423. [101] R.L. Tatusov, E.V. Koonin, D.J. Lipman, Science 278 (1997) 631. [102] R.L. Tatusov, D.A. Natale, I.V. Garkavtsev, T.A. Tatusova, U.T. Shankavaram, B.S. Rao, B. Kirutin, M.Y. Galperin, N.D. Fedorova, E.V. Koonin, Nucleic Acids Res. 29 (2001) 22.